Molecular events in HIV neutralising antibody development

1 September 2014. Related: Basic science.

Gareth Hardy, HIV i-Base

In the May edition of Nature, Nicole Doria-Rose and colleagues, at the Vaccine Research Center, National Institutes of Health, Bethesda, USA, investigated the molecular evolution of an HIV-specific unmutated ancestor antibody through its affinity maturation to an antibody with broadly neutralising capability [1]. These steps may be important guides for the development of a successful HIV vaccine.

Neutralising antibodies against the V1/V2 region of HIV gp120 are the most common cross-reactive neutralising antibodies in natural HIV infection.

They are characterised by long heavy chain complementarity determining region-3 loops (CDR-H3) that protrude, are anionic and are often tyrosine-sulphated. This extended antibody structure is able to penetrate the glycan shield of the HIV envelope protein in order to access its epitope.

Using antibody isolation, B-cell next generation sequencing, structural characterisation and viral single genome amplification Doria-Rose et al delineated longitudinal interactions between the developing antibody and autologous virus in one donor, CAP256, who demonstrated broad virus neutralisation one year after infection. In this individual, superinfection with a second virus was detected 15 weeks after infection with the initial primary virus.

B-cells were isolated from the donor at weeks 59, 119 and 206 after initial infection and used to isolate 12 somatically related monoclonal antibodies, denoted VRC26.01 through to VRC26.12. The same level of virus neutralisation achieved with the donor’s unfractionated plasma was also achieved by use of all 12 mAbs in combination, suggesting this antibody lineage was responsible for the broad and deep neutralising capability of the donor’s plasma.

Using a combination of negative stain electron microscopy, gp120-binding assays and neutralisation fingerprints, the researchers found that the epitope recognised by VRC26 antibodies was similar to that of the PG9 class of neutralising antibodies. This epitope is located at the membrane-distal apex of the gp120 trimer, the specificity of which is dependent on the trimer’s quaternary structure.

Longitudinal sampling of B-cell immunoglobulin sequences with phylogenic analysis revealed that the VRC26 lineage bifurcates from an unmutated common ancestor at about week 38 following infection, giving rise to one branch containing VRC26.01 and one branch containing VRC26.02-12. Thus this data identified the unmutated common ancestor, defined the product of gene recombination in the ancestor B cell and provided a genetic record of the lineage development over the following four years.

Analysis of crystal structures of the antibody Fab fragments of the unmutated common ancestor and six of the other antibodies revealed that the VRC-256 lineage began with an anionic protruding CDR H3, that has structural features similar to other V1V2-specific broadly neutralising antibodies. Over the course of four years almost 20 light chain and more than 30 heavy chain mutations were introduced, including a disulphide bond and the loss of the CDR H3 orientation and its negative charge, although tyrosine sulphation was maintained.

The development of the VRC-256 antibody lineage, together with the selective pressure exerted on viral evolution, were followed by viral single genome amplification (SGA) sequencing over 3 years. Distinct sequences were observed in the V2 region of gp120, which distinguished the primary infecting virus from the superinfecting virus, while substantial recombination between the two viruses had occurred. Before the VRC-26 lineage emerged, most V1V2 sequences in this donor were representative of the primary virus and were neutralisation resistant. While all 12 antibodies effectively neutralised the superinfecting virus, only one, VRC-26.06, neutralised the primary infecting virus. This suggests that the naive B cell that gave rise to the VRC-26 lineage was first engaged by the superinfecting virus.

As the VRC-256 antibody lineage emerged at week 38, a rare K169I mutation occurred in the superinfecting viral sequence that rendered it resistant only to the earliest antibody, VRC-26.01. Therefore VRC-26.01 effectively neutralised the superinfecting virus and drove the selection of mutations that enabled viral escape from this antibody. Once resistance to VRC-26.01 had been achieved, the viral population became predominantly composed of sequences represented by the superinfecting virus. Subsequent somatic mutation of the VRC-26.01 clone gave rise to the development of antibodies (VRC-26.02-12) that neutralised the superinfecting virus. These antibodies corresponded with consistent V1V2 sequences until further viral escape occurred that resulted in a net charge change in the V2 epitope while the VRC-26 CDR-H3s become less acidic. Together the data reveal the co-evolution of viral epitope and antibody specificity, in which the superinfecting virus epitope drove expansion of the VRC-26 lineage.

An effective HIV vaccine should elicit broadly neutralising antibodies. As many neutralising antibodies target the V1V2 region of env, the physical characteristics of the VRC-26 lineage that enable broad neutralisation should be a feature of vaccine-induced V1V2 antibodies: The ability to penetrate the glycan shield and access the V1V2 epitope because they have long CDR H3 regions. Such long CDR H3s only occur in the immunoglobulin VDJ gene rearrangements of an estimated 3.5 – 0.4% of naive B cells, and many of those are auto-reactive and therefore deleted, leaving an even smaller precursor population. This study found that the unmutated common ancestor of the VRC-26 lineage did have a long CDR H3, and that this feature itself did not arise as a result of somatic mutation. Importantly strongly neutralising breadth was achieved from the common ancestor relatively quickly, over a period of months rather than years, resulting from somatic mutation driven by antibody-virus interactions. The authors argue that the crucial factor in developing these antibodies are the engagement of B cells with uncommon receptors that have protruding, anionic, tyrosine-sulphated CDR H3s and that vaccine antigens should be screened that select such B cells.

References:

1. Doria-Rose NA et al. Developmental pathway for potent V1V2-directed HIV-neutralising antibodies. Nature (2014), 509: 55-62.
   http://www.natap.org/2014/HIV/050114_01.htm